Research summary

The purpose of this dissertation is to investigate the positive and negative influences of small-world properties on air transportation network. The methodology and important findings of each part of this dissertation are summarized as follows.

In the first part of this dissertation, we attempted to examine the small-world properties of an international airline alliance network. The effects of the alliance were evaluated by mobility and accessibility models that were formulated based on travel time.

The mobility model was demonstrated to be more appropriate to capture the characteristics of air transportation network than the efficiency model. The results showed that the mobility and accessibility of the post-alliance situation are better than those of the pre-alliance situation, and both situations reveal small-world properties. An alliance can effectively improve the connectivity between airports with high-medium traffic and those with low traffic, and may indirectly induce socioeconomic activities in the latter. After the alliance, some shortest paths may involve more transfers but shorter travel time. The results suggested that airlines should form alliance routes to strengthen the competitiveness of airports with high-medium traffic because alliances can bring

great advantages to such airports and to the airlines involved. That is, airlines can apply our findings to re-design their networks so as to increase their number of passengers and profit. Since the network performance of the three-airlines alliance is better than that of the two-airlines alliance, it suggests that the present alliance can form further alliances with other international airlines to increase the connectivity and efficiency of their networks.

In the second part of this dissertation, we focused on investigating the effects of air travel activities on the transmission of influenza in a small-world air transportation network. We incorporated the air travel activities (passenger consolidation, conveyance and distribution) at airports and on scheduled flights to formulate the transmission models of influenza. For realism of the proposed model, we have weighted each link of the network by the flying time and the number of passengers on the flight. The effects of small-world properties of air transportation network on disease transmission were also discussed. The transmission models were further used to illustrate the spatial and dynamical evolution of disease transmission in the network, and to evaluate the effectiveness of control measures. The results showed that the number or infected individuals and the geographic scope of transmission increased with time, and those airports with a huge number of passengers would accelerate the transmission of disease and cause a much larger pandemic than the average level. This new finding suggests that the public health authorities should avoid the influenza being spread to these airports using complete control strategies. Under the constraint of limited resources, the authorities could in turn practice control measures at airports of different categories, which is an economic and effective approach to mitigate the disease transmission.

However, once the influenza grows up its infectiousness over expectation, the authorities should practice control measures as quickly as possible.

In the third part of this dissertation, we focused on the demand side of air transportation network, in which the first purchase behavior of passengers was investigated in order to illustrate the dynamic adoption pattern of products sold by airlines. Passengers were formed as a small-world social network, i.e. they were not only highly clustered with their friends and acquaintances, but also connected to unfamiliar persons of different social groups by the weak connections. Passengers who obtained the WOM information from their social neighbors would revise their perceptions about products over time, thereby influencing the purchase probability and decision. We incorporated the facts mentioned above into the formulation of the intensity model of WOM information, in which passengers would increase their purchase probability if the intensity of WOM is more than a given threshold. The intensity model of WOM information has never been formulated or discussed in the literature, and may be helpful for relevant airlines and authorities. The results showed that the diffusion pattern of the small-world network lie in between those of regular and random graphs. Without considering the WOM influence in the small-world network, it will result in a mistaken prediction. If the price of LCC is close to that of FSC, then LCC would lose its competitive advantage. However, the LCC could apply a slightly higher price to a lower risk-averse society. These important results provide several managerial implications: (1) applying the same strategy to different characteristic societies may have great risk associated with the failure of marketing; (2) LCCs should pay more attention to the use of WOM communication, and they may release a certain number of free tickets so as to increase the number of passengers with LCC experience, thereby increasing the resulting adopted rate through the transmission of WOM; and (3) LCCs should develop a strategy for enhancing the willingness of passengers to actively or passively share WOM with other people in order to increase the adopted rate.

In the last part of this dissertation, we applied the small-world properties of air transportation network to airport delay problems. The shortcut functions and effects of inter-regional long-haul flights were investigated to illustrate how they cause inter-regional delay propagation across airports and their scheduled flights. The arrival capacity of a given airport was affected by the local weather condition over time, and combining it with different allocation strategies would cause ground holds and delay to flights. We applied an epidemic-like model to illustrate the behavior of delay propagation across airports and flights, in which each airport had two states of disease, whereas each flight had three states of disease. According to the delay duration and the probability of delay occurrence, each flight may incur a certain delays and will not be released until its ground holds is finished. The results showed that according to the average pattern for all airports, a strategy that tends to allocate slots to long-haul flights is the best one among all allocation strategies. However, for those airports that have a high proportion of inter-regional flights scheduled to arrive and provide low capacity to flights due to severe weather, a strategy that is based on the first-scheduled-first-served allocation principle will change to be the best strategy. This kind of strategy is also the best strategy for those airports with a high level of connectivity to other airports. These important and new findings that have never been observed before indicated that, the selection of the best strategy depends on the characteristics of airports, and applying the same strategy to all airports would not result in an optimum condition as one expects before. The results mentioned above may provide helpful insights for the authorities so as to select the best allocation strategy.

The contributions of this dissertation are discussed as follows. This dissertation is the first one to examine the small-world properties of airline alliance networks and to incorporate the air travel activities to construct the models of influence transmission.

This dissertation is also the first one to construct the intensity model of WOM information, and to combine various kinds of social heterogeneity to explore how the adoption behavior of passengers changes over time, and how it is affected by the dynamic properties of network structure. In addition, this dissertation is the first one to combine the small-world properties and an epidemic-like model with the ground-holding policies in order to re-interpret the airport delay problem. This systematical methodology enables us to observe some important results that have never been discussed in the literature. These important results obtained from this dissertation may provide helpful managerial implications for the authorities, airports and airlines when they deal with relevant issues and make decisions.